Optimal experimental designs in RPLC at variable solvent content and pH based on prediction error surfaces

When pH is used as a factor in reversed-phase liquid chromatography (RPLC), it is essential to generate quality, informative data with minimal experimental effort. The most efficient way to achieve this is through experimental designs. Optimising designs that involve solvent content and pH is particularly important, as these factors greatly influence selectivity when ionisable compounds are involved. However, the retention equations for these compounds with pH are nonlinear, making traditional factorial and geometrical designs unsuitable, while D-optimal designs require iterative application. This study explores an extension of G-optimal designs, aimed at improving prediction quality for solvent content and pH. A set of probe ionisable compounds, with well-known retention behaviour, was used. A stepwise strategy provided rapid estimates of the best design with a limited number of experiments. The study focused on determining the optimal distribution and number of points for different acid-base compounds, and the possibility of finding common designs for groups of compounds. Additionally, it aimed to establish design construction rules based on information requirements, without the need for further mathematical treatment.